Inhibitors of HIV Protease (PIs) have proven to be powerful additionsto the armamentariurn for treating HIV infection,but their efficacy has been limited by the emergence of mutantforms of the protease that resist inhibitionby current HPIs. The overall goal of this project is to developand apply new strategies and structure-based computational methods for discovering PIs that will resist existingand, potentially,future mutations. A central element of our approach will be to design compounds based notupon a single protease structure but upon multiple structures. In order to increase the chances of success, combinatorial libraries, rather than singlecompounds, willbe designed and tested. The methodsto be developed will be applicable to other therapeutic targets as well. The specificaims are as follows: 1. Develop one or more scoringfunctionstuned for HIV protease. 2. Use structuralinformationto design combinatoriallibraries of compounds designed to bind and inhibit HIV-1 protease and to evade key current,and potentiallyfuture, resistancemutations. 3. Use detailed, interactivecomputationalmethods to optimizecompounds identified by initialscreening.